Are the experiment files located on a network hard drive or are there copies on each local machine? Are you saving your data to a common network folder? If running over a network or saving data over the network, are you using "unique machine values" to identify the different machines for data writing?